Features
- Added a Max output field to the provider editor, shown next to the Context window, so model reply limits can be set directly in the UI.
- Saving the provider editor now applies the configured output limits to
/v1/models, agent max tokens, and the gateway’s per-member clamp.
Improvements
- The Max output field accepts the same compact format as the Context window, such as
32kfor all models orgpt-6=128k. - Limits not listed in the field are dropped, and models not listed by the provider are refused.
- Added Chinese and Japanese translations for the new field.
Install
Download from usemagpie.ai, or curl -fsSL https://usemagpie.ai/install.sh | sh. An installed magpie updates itself.
- macOS (Apple Silicon):
magpie-darwin-arm64.dmg - macOS (Intel):
magpie-darwin-amd64.dmg - Windows:
magpie-windows-amd64.exe(ARM:magpie-windows-arm64.exe) - Linux:
magpie-linux-amd64/magpie-linux-arm64(needs GTK 3 and WebKitGTK 4.1) - Terminal only:
magpie-cli-<os>-<arch>
功能
- 在提供商编辑器中添加了 最大输出 字段,显示在上下文窗口旁边,因此可以直接在 UI 中设置模型回复上限。
- 保存提供商编辑器现在会将配置的输出上限应用于
/v1/models、代理最大令牌数以及网关的每成员钳制值。
改进
- 最大输出字段接受与上下文窗口相同的紧凑格式,例如
32k用于所有模型,或gpt-6=128k。 - 该字段中未列出的上限将被丢弃,提供商未列出的模型将被拒绝。
- 为新增字段添加了中文和日文翻译。
安装
从 usemagpie.ai 下载,或运行 curl -fsSL https://usemagpie.ai/install.sh | sh。已安装的 magpie 会自动更新。